Forecasting / Input response series data must be non-empty and a column vector.

Hello i am trying to creat a simple forecasting algo for long terms (foraward 10 yrs based on the data from yrs back ) so my code is below, when i run it, i got this error

Input response series data must be non-empty and a column vector.

Error in Untitled2 (line 18)

EstMdl = estimate(Mdl,data);

Please some help guys.

.

.

.

load data

yrs = evolutiondelademand.VarName1;

demand = evolutiondelademand.VarName2;

A= [yrs,demand];

data = iddata(A,[]);

plot (yrs,demand)

%%figure

plot(yrs,demand)

xlabel('years')

ylabel('demand')

past_data = data.OutputData(1:50);

Mdl = arima(2,0,0);

opt = forecastOptions('InitialCondition','e');

K = 100;

EstMdl = estimate(Mdl,data);

[yF,yMSE] = forecast(EstMdl,60,'Y0',data);

legend('Measured','Forecasted')

Answer 1

[Pavl M.]

%load data
%yrs = evolutiondelademand.VarName1;
%demand = evolutiondelademand.VarName2;
odata = xlsread('TitanX Historical Numbers.xlsx');
Dlength = 88;
ts = 1

ts = 1

time = (0:ts:Dlength-1)' %odata(1:Dlength,1)          % Time column (1st column)

time = 88×1

0 1 2 3 4 5 6 7 8 9

<mw-icon class=""></mw-icon>

<mw-icon class=""></mw-icon>

demand = odata(1:Dlength,2)      

demand = 88×1

45499 45500 45501 45502 45503 45504 45505 45506 45507 45508

<mw-icon class=""></mw-icon>

<mw-icon class=""></mw-icon>

%areTimestampsRegular = isregular(demand)
areTimestampsSorted = issorted(demand)

areTimestampsSorted = logical
   1

%w = convert2weekly(odata,Aggregation="mean")
%areTimestampsWRegular = isregular(w,"years")
yrs = double(time)

yrs = 88×1

0 1 2 3 4 5 6 7 8 9

<mw-icon class=""></mw-icon>

<mw-icon class=""></mw-icon>

A= [yrs,demand];
data = iddata(A,[])

data =

Time domain data set with 88 samples.
Sample time: 1 seconds                
                                      
Outputs      Unit (if specified)      
   y1                                 
   y2                                 
                                      

figure
plot(yrs,demand)
xlabel('years')
ylabel('demand')
title('InputData')

Npastsamples = 50;
past_data = data.OutputData(1:Npastsamples);
%nMdl0 = nlarx(data, [2 2 1;2 2 1])
Mdl0 = arima(3,2,1)

Mdl0 = 
  arima with properties:

     Description: "ARIMA(3,2,1) Model (Gaussian Distribution)"
      SeriesName: "Y"
    Distribution: Name = "Gaussian"
               P: 5
               D: 2
               Q: 1
        Constant: NaN
              AR: {NaN NaN NaN} at lags [1 2 3]
             SAR: {}
              MA: {NaN} at lag [1]
             SMA: {}
     Seasonality: 0
            Beta: [1×0]
        Variance: NaN

Mdl = gjr(9,21)

Mdl = 
  gjr with properties:

     Description: "GJR(9,21) Conditional Variance Model (Gaussian Distribution)"
      SeriesName: "Y"
    Distribution: Name = "Gaussian"
               P: 9
               Q: 21
        Constant: NaN
           GARCH: {NaN NaN NaN NaN NaN NaN NaN NaN NaN} at lags [1 2 3 4 5 6 7 8 9]
            ARCH: {NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN} at lags [1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21]
        Leverage: {NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN} at lags [1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21]
          Offset: 0

Mdl2 = garch(Constant=0.05,GARCH=0.91,ARCH=0.04) %Stationary Model

Mdl2 = 
  garch with properties:

     Description: "GARCH(1,1) Conditional Variance Model (Gaussian Distribution)"
      SeriesName: "Y"
    Distribution: Name = "Gaussian"
               P: 1
               Q: 1
        Constant: 0.05
           GARCH: {0.91} at lag [1]
            ARCH: {0.04} at lag [1]
          Offset: 0

opt = forecastOptions('InitialCondition','e');
K = 100;
Nhor = 60; %actual forecasting horizon

rng("default") % For reproducibility

[vS,yS] = simulate(Mdl2,50)

vS = 50×1

0.9716 1.0648 1.2363 1.2118 1.1577 1.1827 1.1351 1.0883 1.5978 1.9941

<mw-icon class=""></mw-icon>

<mw-icon class=""></mw-icon>

yS = 50×1

1.8076 -2.3309 0.9586 0.3509 -1.4070 -0.4715 0.3650 3.7330 3.5006 -1.9062

<mw-icon class=""></mw-icon>

<mw-icon class=""></mw-icon>

figure
plot(vS)

figure
plot(yS)

y0 = yS(1);
v0 = vS(1);
y = yS(2:end);
v = vS(2:end);

res = infer(Mdl2,y) % Retrieve inferred residuals

res = 49×1

2.4242 2.4734 2.3375 2.1821 2.1149 1.9834 1.8602 2.3002 2.6334 2.5917

<mw-icon class=""></mw-icon>

<mw-icon class=""></mw-icon>

figure
plot(res)

foreValues2 = forecast(Mdl2,Npastsamples,'Y0',demand) % forecast

foreValues2 = 50×1

1.0e+08 * 9.2344 8.7727 8.3341 7.9174 7.5215 7.1454 6.7882 6.4488 6.1263 5.8200

<mw-icon class=""></mw-icon>

<mw-icon class=""></mw-icon>

figure
plot(foreValues2)

data(:,1)'

ans =

Time domain data set with 88 samples.
Sample time: 1 seconds                
                                      
Outputs      Unit (if specified)      
   y1                                 
                                      

EstMdl = estimate(Mdl0,demand);

Warning: Lower bound constraints are active; standard errors may be inaccurate.

 
    ARIMA(3,2,1) Model (Gaussian Distribution):
 
                Value    StandardError    TStatistic    PValue 
                _____    _____________    __________    _______

    Constant        0              0           NaN          NaN
    AR{1}           0              0           NaN          NaN
    AR{2}           0              0           NaN          NaN
    AR{3}       2e-12              0           Inf            0
    MA{1}       2e-12              0           Inf            0
    Variance    2e-07     2.6752e-07       0.74759      0.45471

[yF,MSEError] = forecast(EstMdl,Nhor,'Y0',demand)

yF = 60×1

45587 45588 45589 45590 45591 45592 45593 45594 45595 45596

<mw-icon class=""></mw-icon>

<mw-icon class=""></mw-icon>

MSEError = 60×1

1.0e+00 * 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0001 0.0001

<mw-icon class=""></mw-icon>

<mw-icon class=""></mw-icon>

tn = [yrs; (yrs(end):1:yrs(end)+Nhor-1)']

tn = 148×1

0 1 2 3 4 5 6 7 8 9

<mw-icon class=""></mw-icon>

<mw-icon class=""></mw-icon>

figure
plot(tn,[demand; zeros(1,Nhor)'],'b',tn,[zeros(1,Dlength)'; yF],'g')
xlabel('years')
ylabel('demand')
title('Historical in blue and Forecast in green')

figure
plot(tn,[demand; yF])
xlabel('years')
ylabel('demand')
title('Merged Historical and Forecast Data')

%Green curve after input max year Dlength=88 is predicted(forseen,anticipated), 
% so the predictions/forecast are correct
% correct )))